Vitamin D Supplementation To Prevent Acute Respiratory Infections: Individual Participant Data Meta-analysis
Notes for this study:
||Event rate per participant-year. Control: 121/284.7 (0.43) Vitamin D: 85/286.6 (0.30) Adjusted incidence rate ratio 0.74 (0.56 to 0.97)
|Number of Subjects
||0-1, 1-6, 7-12, 13-17, 18-29, 30-44, 45-64, 65+
An individual participant data meta-analysis was performed with 25 eligible studies and 11,321 participants aged from 0-95. The major outcome of interest was the effect of vitamin D supplementation on the risk of acute respiratory infections (ARIs).
The eligibility criteria were randomized, double-blind, placebo-controlled trials that used vitamin D, of any duration, and analyzed acute respiratory infections. The meta-analysis was a random-effects model adjusted for age, sex, and study duration, and the outcomes specifically were the proportion of participants experiencing at least one ARI, ARI rate, and time to first ARI.
Overall, the proportion of participants with at least one ARI was modestly but statistically significantly lower for vitamin D when adjusted. For those with vitamin D levels of less than 25 nmol/L, this was considerably more potent than those with higher levels, and only statistically significant for those with low levels. Bolus doses were significantly less effective than non-Bolus doses, though there was a smaller statistically significant reduction for those with higher levels of vitamin D for daily/weekly supplementation, but not for bolus doses. The event rate was slightly but significantly lower overall in the vitamin D groups, and more notably, lower with low vitamin D levels, whereas there wasn't an effect bolus dose. The time until the first infection was nonsignificantly longer when in the vitamin D group, with no significant effect of baseline status or bolus dose.
A protective effect against asthma exacerbation was found, but not for chronic obstructive pulmonary disease.
There wasn't a significantly different rate of serious adverse events or mortality between groups.
Overall, the studies were of low risk for bias.